They conducted a take home coding test and subsequent interview rounds in which they tested ML concepts. There was also a coding round with leetcode questions on linked list operations.

Had 3 hours of interview, starting with 30 minutes presentation about previous work and research. Second hour covered ml system design, third hour covered paper reading and implementation with live coding session

The interview began with LeetCode-style coding problems to assess algorithmic fundamentals, followed by in-depth discussions on AI infrastructure, covering distributed training, inference serving, data pipelines, resource scheduling, and system design for large-scale machine learning systems.